Living cells use molecular control circuits to achieve complex behaviors such as homeostasis and adaptation. Typically, such systems are studied on a case-by-case basis, using reverse engineering to dissect the structure of individual regulatory networks. Because of the obscuring effects of evolution, such approaches often fail to reveal the core design principles underlying function. To overcome this problem we propose to apply forward engineering and comparative genomic approaches to understand the design principles of biological circuits. We will focus on understanding adaptation - a property of many sensory systems in which cells respond to an input in a transient manner, but then reset to allow response to further increases in input. We will ask not how any one particular system works, but rather what are the core network structures required to achieve adaptation, and what are the diverse ways in which such structures can be implemented with biological components. Our specific aims are to: (1) Theoretically define the design rules of adaptation circuits - computationally enumerate and classify all core architecture families that can robustly perform adaptation. (2) Build synthetic adaptation circuits - empirically test and refine our models for circuit structure/function relationships. These will be constructed in yeast using a toolkit of modular molecular parts optimized for engineering of kinase networks. (3) Analyze the variation in natural adaptation circuits across species-- we will focus primarily on the osmostress response in diverse fungal species, but will also quantitatively explore adaptation circuits in other yeast and mammalian stress response pathways.